Party support over time
About
This tool tracks public opinion polling for Poland's next parliamentary election, showing how party support shifts over time. Instead of looking at individual polls in isolation, which can be noisy and vary by polling agency, it calculates rolling averages over a configurable time window. This way, you can smooth out the noise and see underlying trends more clearly.
You control how the averages are built: choose your window size, weight recent polls more heavily, decide how to handle undecided voters, and pick which agencies to include or emphasize. Everything runs entirely in your browser: no data is sent to any server, and you can load your own poll data in CSV format, as well as download source data.
Frequently asked questions
According to default assumptions on 2026-10-11: KO 31.05%, PiS 19.58%, Konfederacja 13.64%, KKP 8.34%, Lewica 7.44%, R+ 5.07%, Razem 4.17%, PSL 2.89%, Polska 2050 1.15%. Polls taken into account: IBRiS (United Surveys) β 3, Opinia24 β 2, Social Changes β 2, Pollster β 1, OGB β 1, Research Partner β 1, IBRiS β 1, CBOS β 1. The figures update whenever new polls are added to the dataset. You can change assumptions to see how the numbers shift.
How is the average calculated?Each point on the chart is the average of all polls whose date falls within the last N days (35 by default). You can choose between a weighted mean and a median, give more weight to recent polls using exponential decay (the default half-life is 14 days) or polls from more credible companies, and decide whether each poll is dated by its start, midpoint, or end date. A shaded band can show the minimum and maximum for each day.
Where does the poll data come from, and which pollsters are included?The data comes from the English Wikipedia article on opinion polling for the next Polish parliamentary election. Pollster names are used only to identify the source of each published poll; this site is not affiliated with any pollster or party. You can include, exclude, or reweight individual pollsters in the settings.
How are undecided voters handled?Undecided or "don't know" responses can be shown as a separate category, split proportionally among the parties shown, or split according to percentages you set yourself. Different pollsters report undecided voters differently.
Why do polls from different companies differ?Pollsters differ in sample size, interview mode (phone, online, or mixed), question wording, weighting methods, and how they treat undecided respondents. Some also show a consistent lean relative to the average, often called a house effect. Averaging several polls smooths out random noise and some of these differences, but it can't remove errors shared by all pollsters.
Is a poll average a prediction of the election result?No. It summarizes published polls at a given point in time. Opinion can change before an election, polls carry sampling error, and assumptions may be erroneous. The latter is why polls.nawa.ro allows setting them according to your own preferences.